There are multiple ways to do this, depending on how you're performing the cluster upgrade.
option A, headswap keeping the same storage.
option B, add the new A1 nodes to the existing A cluster and perform vol moves to the aggregates on the A1 storage before decommissioning (cluster unjoin) the old A cluster nodes. Then, follow the same action plan for the B/B1 cluster.
option C, SnapMirror to the new A1/B1 clusters via fan out mirrors, this should still have a common snapshot.
there are several possible issues with resync not being able to find common snapshots, below are a few examples, but I'd be interested in the exact failure you're getting.